2018
Publication Year
2021-01-13T15:27:51Z
Acceptance in OA@INAF
The Carnegie-Chicago Hubble Program: Calibration of the Near-infrared RR Lyrae
Period-Luminosity Relation with HST
Title
Rich, Jeffrey A.; Madore, Barry F.; Monson, Andrew J.; Freedman, Wendy L.;
Beaton, Rachael L.; et al.
Authors
10.3847/1538-4357/aae980
DOI
http://hdl.handle.net/20.500.12386/29754
Handle
THE ASTROPHYSICAL JOURNAL
Journal
869
The Carnegie-Chicago Hubble Program: Calibration of the Near-infrared RR Lyrae
Period
–Luminosity Relation with HST
Jeffrey A. Rich1 , Barry F. Madore1 , Andrew J. Monson2, Wendy L. Freedman3, Rachael L. Beaton4,11,12 , Gisella Clementini5, Alessia Garofalo6,7, Dylan Hatt3 , Taylor Hoyt3 , In-Sung Jang8 , Juna A. Kollmeier1 ,
Myung Gyoon Lee8 , Jillian R. Neeley9 , Victoria Scowcroft10 , and Mark Seibert1
1
The Observatories of the Carnegie Institution for Science, 813 Santa Barbara Street, Pasadena, CA 91101, USA;jrich@carnegiescience.edu
2
Department of Astronomy and Astrophysics, The Pennsylvania State University, 525 Davey Lab, University Park, PA 16802, USA
3
Department of Astronomy and Astrophysics, University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637, USA
4
Department of Astrophysical Sciences, Princeton University, 4 Ivy Lane, Princeton, NJ 08544, USA
5
INAF-Osservatorio di Astrofisica e Scienza dello Spazio, via Piero Gobetti 93/3, I-40129, Bologna, Italy
6
INAF-Osservatorio di Astrofisica e Scienza dello Spazio, via Piero Gobetti 93/2, I-40129, Bologna, Italy
7
Dipartimento di Fisica e Astronomia, Universitá di Bologna, Viale Berti Pichat 6/2, I-40127 Bologna, Italy
8
Department of Physics & Astronomy, Seoul National University, Gwanak-gu, Seoul 151-742, Republic of Korea
9
University of Florida, Department of Astronomy, 211 Bryant Space Science Center P.O. 112055, Gainesville, FL, USA
10
Department of Physics, University of Bath, Claverton Down, Bath, BA2 7AY, UK
Received 2018 March 6; revised 2018 October 10; accepted 2018 October 10; published 2018 December 13
Abstract
We present photometry of 30 Galactic RR Lyrae variables taken with HST WFC3/IR for the Carnegie-Chicago Hubble Program. These measurements form the base of the distance-ladder measurements that comprise a pure Population II base to a measurement of Hoat an accuracy of 3%. These data are taken with the same instrument and filter (F160W) as our observations of RR Lyrae stars in external galaxies so as to minimize sources of systematic error in our calibration of the extragalactic distance scale. We calculate mean magnitudes based on one to three measurements for each RR Lyrae star using star-by-star templates generated from densely time-sampled data at optical and midinfrared wavelengths. We use four RR Lyrae stars from our sample with well-measured HST parallaxes to determine a zero-point. This zero-point will soon be improved with the large number of precise parallaxes to be provided by Gaia. We also provide preliminary calibration with the TGAS and Gaia DR2 data, and all three zero points are in agreement, to within their uncertainties.
Key words: stars: distances – stars: variables: RR Lyrae Supporting material: machine-readable tables
1. Introduction
Since the discovery of the expanding universe by Hubble (1929) astronomers have measured the rate of expansion, the
Hubble Constant(Ho), with increasing accuracy. The comple-tion of the Hubble Space Telescope (HST) Key Project established the modern era of Ho measurements, with more recent Cepheid-based distance-ladder studies reaching the 3% level of precision (Freedman et al. 2001, 2012; Riess et al.
2016). The Carnegie-Chicago Hubble Program (CCHP) aims
to provide a new and independent pathway to the measurement of Ho. Using a distance ladder comprised of RR Lyrae stars, Tip of the Red Giant Branch (TRGB) stars, and Type Ia Supernovae(SNe Ia), CCHP-II will provide a pure Population II basis to a measurement of Ho, with an aim of achieving an accuracy of 3% (Beaton et al. 2016, W. L. Freedman et al. 2018, in preparation). In this way, galaxies with only old stellar populations(and thus no Cepheid variables) can also be used as an independent assessment of the ongoing tension between distance-ladder-based and cosmic-microwave-background-based measurements of Ho (Freedman 2017; Riess et al.
2018). The SNe Ia zero-point can be set using only the TRGB
rung of the distance ladder, with both current TRGB parallax measurements (e.g., Brown et al.2018) and in the future with
Gaia parallaxes. The TRGB zero-point itself, however, can be
independently set by the RR Lyrae stars. Thus thefirst step in this distance ladder will be enabled by a well-calibrated, accurate zero-point for Galactic RR Lyrae stars.
The existence of a near-infrared PL relation for RR Lyrae stars was first demonstrated by Longmore et al. (1986, 1990)
with K-band observations of RR Lyrae stars in Galactic globular clusters. Theoretical models now also clearly show that a PL relation for RR Lyrae stars is to be expected at nearly all wavelengths, with the exception of the V-Band(e.g., Bono et al. 2003; Catelan et al. 2004). Furthermore, Madore &
Freedman (2012) demonstrated that there is a decrease in
intrinsic scatter with an increase in wavelength of PL relations for both Cepheids and RR Lyrae stars due to a number of effects including a lower sensitivity of surface brightness to temperature variations and, to a lesser extent, a decrease in the line-of-sight extinction. As such, IR-wavelength determinations of the PL relation for RR Lyrae stars have significant benefits when compared to visible light measurements.
RR Lyrae PL relations are well characterized from optical (R and I) to mid-IR wavelengths (e.g., Madore et al. 2013; Dambis et al.2014; Braga et al.2015; Neeley et al.2015), but in
order to use RR Lyrae stars as part of a precision distance ladder in the determination of Ho, it is essential to reduce both random and systematic errors(Freedman et al.2012; Beaton et al.2016; Riess et al.2016). As such, the CCHP was designed to employ
the smallest number of instruments and filter combinations possible across the entire distance ladder. To minimize
© 2018. The American Astronomical Society. All rights reserved.
11
Hubble Fellow.
12
systematics, the CCHP Hocalibration, in fact, seeks a purely space-based calibration of Howith the use of a single telescope, the HST. This is accomplished by calibrating the RR Lyrae PL relation at the reddest near-IR wavelengths available to HST, with the F160W (H-band) filter on the the WFC3/IR camera.
A near-IR calibration of the Galactic RR Lyrae PL zero-point is now possible given the availability of accurate(∼8%) parallaxes determined for a few stars using the HST Fine Guidance Sensor(FGS; Benedict et al. 2011) and in the near
future Gaia will provide parallaxes at similar or better precision for hundreds of RR Lyrae stars (Clementini 2016; Gaia Collaboration et al. 2017). In this paper, we present our
measurement of the geometrically based period–luminosity relation for Galactic RR Lyrae stars with HST WFC3/IR photometry that can now be used to calibrate this particularfirst rung of the CCHP-II distance ladder.
2. Sample and Observations
Our sample consists of 30 Galactic RR Lyrae stars. Five of our RR Lyrae stars have trigonometric parallaxes measured with HST/FGS: XZ Cyg, UV Oct, RR Lyr, SU Dra, and RZ Cep (Benedict et al. 2011). The remaining 25 Galactic RR
Lyrae stars will ultimately have their parallaxes determined by Gaia with an error of ∼1% (de Bruijne et al. 2014; Clementini 2016; Gaia Collaboration et al. 2017). Our
observations were conducted with HST WFC3/IR and were spread across two HST programs as described in detail in the following subsections. The first program consisted of single-epoch SNAP observations designed to form the foundation of our full HST-based calibration(PID 13472, PI Freedman). The second program was designed to link the RR Lyrae stars and TRGB calibrations out to the Hubble flow, with dedicated follow-up and single-epoch observations for thefive RR Lyrae stars with accurately measured parallaxes (PID 13691, PI Freedman). The final sample is presented in Table1, periods are adopted from Monson et al. (2017), wherein a
self-consistent solution for multiwavelength, long time-baseline data was found for each star in our sample. All of our stars have B, V, I, and Spitzer 3.6 and 4.5μm data.
2.1. SNAP Observing Program
The majority of our observations were taken in the course of our HST SNAP program. Each star observed in the SNAP program has only one randomly timed epoch due to the nature of SNAP observations. With the exception of RR Lyrae itself, our data consist of several very short direct exposures due to the brightness of our targets and the desire to gather as much data as possible in the portion of each HST orbit we were allotted.
For every star it was necessary to use both the subarray and RAPID sampling readout modes to avoid saturating the detector. Based on saturation times estimated from the WFC3/IR exposure time calculator (ETC), stars were observed with either a 64×64 or 128×128 subarray with a single RAPID sample, resulting in the shortest possible total exposure times of 0.061 and 0.113 s respectively. These rapid observa-tions are further facilitated by the lack of a shutter for the WFC3/IR channel. This was done 26 times in sequence with a simple two-point dither pattern, achieving the maximum number of exposures possible before a readout of the WFC3/ IR buffer was necessary. The very short exposure and subarray
size ensured that in all frames the only object with any detectableflux was the RR Lyrae variable targeted.
Based on the previous observations of bright Cepheids (Riess et al.2014and Casertano et al.2016), we observed the
star RR Lyr using WFC3/IR in a drift scanning mode. In order to drift rapidly enough that RR Lyr did not saturate, our observations were carried out with a full-frame readout in gyro-only guiding mode, facilitating a scan rate of 7.5 arcsec s−1. With the maximum number of RAPID samples possible as the star sped across the detector, our total time spent on target for a single drift scan of RR Lyrae was 17.6 s. This observation was repeated four times with a two-point dither pattern, with the number of exposures again limited by the maximum number of full-frame observations possible before a buffer read was required.
2.2. GO Observing Program
The remaining observations of our sample were taken as part of our large HST General Observing program summarized in Beaton et al.(2016). We used WFC3/IR to observe the five
Galactic RR Lyrae stars from Benedict et al.(2011). This was
done to achieve two sets of measurements separated in phase in order to better constrain the mean magnitudes of our zero-point calibrators and to allow us to assess the accuracy of our single-phase SNAP-based magnitudes.
We found that despite the predictions of the ETC, even the shortest exposures in our SNAP observations with WFC3/IR show evidence of nonlinearity. Because of this, our observa-tions included drift scanning observaobserva-tions to avoid reaching nonlinearity in an effort to cross-check our short-exposure SNAP and GO observations. We split a single orbit between direct and drift scanning observations. One star, UV-Oct, was allocated two orbits as it had not yet been observed in our SNAP program, while the remaining four stars were allocated one orbit each.
As with the SNAP observations, we observed each of the five targets with the smallest 64×64 subarray and RAPID sampling to minimize the number of pixels subject to nonlinearity, resulting in 0.061 second individual exposures. We used a four-point dither pattern for these observations with eight cycles over a single orbit, resulting in 32 individual exposures of each star.
After the direct imaging exposures, we again used the gyro-only guiding mode to observe in drift scanning mode with full-frame readout. Each star was observed in this manner with four exposures on different portions of the chip, with a total exposure time of 14.66 s for each, with a scan rate of 7.5 arcsec s−1for UV Oct and 6.5 arcsec s−1for the remaining four stars.
3. Analysis
Although none of our direct imaging observations contain truly saturated pixels(with the exception of RR Lyr), some of our observations have pixels that surpass the nonlinearity threshold established in the HST WFC3/IR manual and are likely approaching saturation between the zeroth andfirst (and only) read, resulting in an estimation of counts from the zeroth read only and a significant uncertainty in the total number of counts(Riess2011). For these frames the estimation of counts
is significantly overestimated, and the nonlinearity correction applied by the HST pipeline to the calibrated,flat-fielded data
results in significant deviations from expected values. That is, simple aperture photometry of observations with these significantly brighter pixels results in highly inaccurate measurements and suggests that the actualflux is overpredicted in such cases.
Figure1shows each of the direct images of XZ Cyg obtained in the SNAP program; 12 of the exposures do not approach the “saturation” threshold (blue dots), while 14 of the exposures have one pixel with a significant nonlinearity correction applied (red dots). Aperture photometry of the individual HST-calibrated FLT files shows a a bimodal pattern with a systematic difference of 0.2 mag. The bimodality is a result of the dither pattern: in the exposures with a correction applied, the PSF of the star is centered perfectly on a single pixel while in the other exposures the center falls in a gap between pixels, lowering the peakflux in any one pixel below the nonlinearity threshold.
Although the stars are well isolated, the presence of significantly nonlinear pixels without a proper correction means that we cannot apply simple aperture photometry to our direct SNAP and GO observations. Our solution to this situation is to mask pixels in frames with poor nonlinearity corrections and then undertake PSF photometry.
Table 1 Galactic RR Lyrae Sample
Star Period(days) Observations Type Blazhko [Fe/H] AF160W
SW And 0.442270 a RRab * −0.24 0.021 UY Cam 0.267042 a RRc L −1.33 0.012 RZ Cep 0.308710 a,c,d RRc L −1.77 0.043 RR Cet 0.553028 a RRab L −1.45 0.012 XZ Cyg 0.466599 a, c, d RRab * −1.44 0.053 DX Del 0.472617 a RRab L −0.40 0.051 SU Dra 0.660420 a, c, d RRab −1.80 0.006 CS Eri 0.311331 a RRc L −1.41 0.010 SV Eri 0.713796 a RRab L −1.70 0.047 RR Gem 0.397290 a RRab * −0.29 0.030 TW Her 0.399600 a RRab L −0.69 0.023 BX Leo 0.362860 a RRc L −1.28 0.013 RR Leo 0.452393 a RRab L −1.60 0.020 TT Lyn 0.597434 a RRab L −1.56 0.009 RR Lyr 0.566838 a, b, c, d RRab * −1.39 0.017 UV Oct 0.542580 a, c, d RRab * −1.74 0.050 BH Peg 0.640993 a RRab * −1.22 0.043 RU Psc 0.390385 a RRc * −1.75 0.024 HK Pup 0.734238 a RRab L −1.11 0.088 RU Scl 0.493339 a RRab L −1.27 0.010 SV Scl 0.377340 a RRc L −1.77 0.008 AN Ser 0.522072 a RRab L −0.07 0.022 V440 Sgr 0.477479 a RRab L −1.40 0.047 V675 Sgr 0.642289 a RRab L −2.28 0.072 MT Tel 0.316900 a RRc L −1.85 0.021 AM Tuc 0.405810 a RRc * −1.49 0.013 AB UMa 0.599577 a RRab L −0.49 0.012 RV UMa 0.468060 a RRab * −1.20 0.010 SX UMa 0.307118 a RRc L −1.81 0.006 TU UMa 0.557659 a RRab L −1.51 0.012
Note.(a) Direct imaging in SNAP program #13472. (b) Drift-scan imaging in SNAP program #13472. (c) Direct imaging in GO program #13691. (d) Drift-scan imaging in GO program#13691.
(This table is available in machine-readable form.)
Figure 1.Aperture photometry of the SNAP observations of XZ Cyg. Each point corresponds to a single frame, points are color-coded based on their dither position. The∼0.2 mag difference caused by the dither pattern is a result of an overcorrection for nonlinearity in a single bright pixel at dither point a.
3.1. Direct Imaging Photometry
Photometry was carried out with DAOPHOT (Stet-son1987,1994) on all individual direct imaging “FLT” frames
with the HST Pixel Area Map correction applied. Because we cannot generate a unique PSF for each frame, we used DAOPHOT to generate and apply one PSF for use in the 64×64 images and then generated a separate PSF for the 128×128 images. To generate the PSF, we used stars that are “clean” (i.e., images that do not have pixels that approach the nonlinear regime): 156 images for the 64×64 subarray and 208 for the 128×128 subarray or 26 images of six and eight stars each, respectively.
To determine which parameters to assign to our model PSF, we generated several thousand individual PSF models with PSF radius andfitting radius varied in integer pixel increments from 2 to 17 and the seven different PSF functional forms available in DAOPHOT. We then compared the magnitude derived from each model PSF for each star with aperture photometry and chose PSF parameters that, on average, best reproduced the aperture magnitudes. We assigned a PSF radius of 14 pixels to the 64×64 PSF model and 17 pixels for the 128×128 PSF model and used a fitting radius of 12 pixels and a Moffat function withβ=3.5 for both PSF models.
We also use the “clean” images to test the effect of pixel masking on our PSF photometry. We chose the brightest pixel in each image, masked it, and redetermined the magnitude using the same range of PSF parameter values. Regardless of the parameters used, masking pixels did not significantly impact DAOPHOT PSF photometry measurements. When masking stars in our 64×64 images, we found a systematic offset of −0.0017±0.0022 mag, and for 128×128 images we found an offset of−0.0024±0.0028 mag.
Finally, we calculated aperture corrections to use with our PSF photometry with DAOGROW and aperture photometry of our “clean” images (Stetson 1990). We applied an aperture
correction of 0.0196 mag to 64× 64 images with a 14 pixel radius PSF and 0.0136 mag to 128× 128 images with a 17 pixel PSF. These aperture corrections were computed to an infinite aperture, after which we applied the HST WFC3/IR F160W Vega zero-point for an infinite aperture of 24.695 mag (taken fromhttp://www.stsci.edu/hst/wfc3/ir_phot_zpt). The
final photometry for each observation, for each target, is given in Table 2. The errors quoted in Table 2 are the photometric errors derived by DAOPHOT. The mean magnitudes derived in Section 2 include an additional 2% systematic error as described athttp://stsci.edu/hst/wfc3/phot_zp_lbn.
4. Mean Magnitude Determination
Because we have only sparsely sampled light curves, we have undertaken templatefitting to determine mean magnitudes for our sample of RR Lyrae stars. Fortunately, the stars in our sample have a wealth of ground- and space-based observations available for the construction of light-curve templates con-structed on a star-by-star basis (Monson et al. 2017). We
consider two approaches: (1) we compare our observations to smooth light curves constructed from Spitzer 3.6μm observa-tions and(2) we use predicted templates generated from optical V and I band ground-based observations, asfirst proposed and demonstrated for Cepheids in Freedman & Madore(2010a). In
each case, we averaged all of the data points available in each epoch of SNAP or GO observation into a single-epoch, intensity-averaged point, resulting in one to three HST F160W points to which a template light curve can be compared. Phases for the data points used in mean magnitude determination are calculated using the periods and phase correction information and methodology to be outlined in R. L. Beaton et al.(2018, in preparation).
The smoothed Spitzer light curves were generated by applying GLOESS (Gaussian-windowed LOcal regrESSion method, Persson et al. 2004; Monson et al. 2017) to 20–40
evenly spaced data points, resulting in a smoothly varying curve(Monson et al.2017). The mid-IR data are well suited for
comparison to our near-IR observations, as we expect by the H-band that the light-curve shape is dominated by radius effects. As such, we simply offset in magnitude the 3.6μm GLOESS curve tofit the available HST data points and take the mean of the offset light curve to be our F160W mean magnitude. Figure2demonstrates this process for UV Oct, the star in our sample with the most data points. The error on our mean magnitude measurement combines the random
Table 2
PSF Photometry of Individual Exposures
Name MJDobs Phase F160W F160W
(mag) (σ)
AM Tuc 56900.43533022 0.98375 10.677 0.030
AM Tuc 56900.43553856 0.98426 10.676 0.030
AM Tuc 56900.43574688 0.98477 10.651 0.030
AM Tuc 56900.43642986 0.98646 10.699 0.030
Note.Example of photometry of individual exposures. (This table is available in its entirety in machine-readable form.)
Figure 2.Example of our HST PSF photometry(blue points) for one RR Lyrae variable(UV Oct) compared to Spitzer data and the corresponding GLOESS light curve(yellow squares and line). The plot shows the GLOESS fit to the Spitzer data prior to scaling(above) and after scaling (below) to the F160W data. The mean magnitude derived from the shifted curve is shown as a horizontal line.
photometric error of our data, the error in the template, and the error on the phase estimate and its location with respect to the template. There is an additional possible systematic error induced by the difference in amplitude between 3.6μm and the H-band. Measurements by Monson et al.(2017) show that the
H-band has a larger amplitude with a median of ∼0.03 magnitudes, leading to a potential offset in magnitude of∼0.01 magnitudes, depending on the phase of our measurements.
For comparison with our Spitzer-derived magnitudes, we follow the same procedure with custom-generated template light curves for each RR Lyrae variable. We use near-infrared templates generated from V and I band data using a process described in Beaton et al. (2016), Monson et al. (2017), and
R. L. Beaton et al.(2018, in preparation). The templates require well-sampled B, V and I light curves that can be combined to trace the change in temperature and radius. Temperature-independent curves that predict the shape of the IR light curves are generated using the visible light. These curves predict the shape of the IR light curves that primarily trace radius, with a scaling in amplitude (Freedman & Madore2010a,2010b).
We offset the template to match the available data and measure the mean of the shifted ground-based template. The intensity-averaged measured mean magnitudes for both methods are given in Table 3. The two methods provide consistent mean magnitudes, with an average difference of
0.005±0.019 mag (magSptz–magtmpl). Furthermore, our F160W magnitudes are offset by 0.015±0.06 mag from the Two Micron All-Sky Survey (2MASS) H-band magnitudes, which is roughly consistent with the 0.0215±0.0054 mag difference calculated by Riess(2011).
5. Period–Luminosity Relations
To construct our PL relations we calculate absolute magnitudes for RR Lyrae stars from our F160W sample with known parallaxes. For this calculation we use the mean magnitudes derived by scaling Spitzer data as described in the previous section. We examine RR Lyrae stars with HST/FGS parallaxes (Benedict et al. 2011), Tycho-Gaia Astrometric
Solution (TGAS) parallaxes (Lindegren et al. 2016; Gaia Collaboration et al.2017; Gould & Kollmeier2017), and Gaia
data release 2 (DR2) parallaxes (Gaia Collaboration et al.
2018; Lindegren et al.2018). Our absolute magnitudes include
a correction for extinction adopted from the published values of Feast et al.(2008) that will be discussed further in R. L. Beaton
et al.(2018, in preparation). The values we use in this paper are given in Table1. As a check for consistency, we compare our absolute magnitudes and PL relations with 2MASS H-band magnitudes. We consider only the fundamental pulsators (RRab) in our PL analysis.
To compute the absolute magnitude, we use two methods: we invert the parallax to produce a distance and we use the astrometry-based luminosity(ABL) method for parallaxes with larger errors. For direct inversion of parallaxes, we restrict our sample toσπ/π<0.1, the regime where this is an appropriate treatment for the resulting distance uncertainties and no additional forward modeling is explicitly required(Astraatmadja & Bailer-Jones2016; Gaia Collaboration et al.2017). We use
only RRab stars for determination of the zero points.13
Absolute magnitudes derived from directly inverted paral-laxes used in our PL analysis are given in Table4. The errors for absolute magnitudes derived with HST/FGS and TGAS parallaxes are calculated directly from the errors in the parallax as they dominate the uncertainty. The error in absolute magnitude when derived from Gaia DR2 parallaxes are a combination of the error on the mean magnitude(Table3) and
the error in parallax. The period–luminosity relation takes the simple form:
Ml=al´[log( )P -0.3]+bl, ( )1 where a and b are the slope and intercept, respectively, and P is the period in days. We perform an unweighted least-squares (LSQ) fit assuming a fixed slope of a=−2.215 and a=−2.189 for fundamental pulsators in the 2MASS H-band (Braga et al.2015) and HST F160W respectively (M. Marconi
et al. 2018, in preparation). Using the cuts discussed in the previous section, four stars with HST/FGS parallaxes, three with TGAS parallaxes, and 20 stars with DR2 parallaxes are used to generate different PL relations this way. The resulting zero points at log(P)=−0.3 days are given in Table5.
Example PL relationfits are shown in Figure3.
Table 3 Mean Magnitudes
Name F160W F160W F160W F160W 2MASS
Spitzer σs Template σT H
(mag) (mag) (mag) (mag) (mag)
SW And 8.606 0.020 8.598 0.020 8.590 UY Cam 10.851 0.024 10.849 0.025 10.841 RZ Cep 7.972 0.016 8.007 0.017 8.136 RR Cet 8.609 0.025 8.617 0.023 8.652 XZ Cyg 8.665 0.020 8.749 0.025 8.770 DX Del 8.752 0.027 8.750 0.064 8.818 SU Dra 8.701 0.018 8.733 0.019 8.686 CS Eri 8.174 0.024 8.217 0.025 8.175 SV Eri 8.657 0.023 8.718 0.024 8.736 RR Gem 10.334 0.023 10.339 0.021 10.271 TW Her 10.385 0.041 10.352 0.027 10.269 BX Leo 10.688 0.034 10.750 0.058 10.743 RR Leo 9.672 0.021 9.798 0.028 9.737 TT Lyn 8.697 0.020 8.701 0.022 8.656 RR Lyr 6.472 0.035 6.481 0.031 6.598 UV Oct 8.316 0.013 8.292 0.013 8.289 BH Peg 9.180 0.025 9.184 0.029 9.070 RU Psc 9.190 0.025 9.244 0.021 9.156 HK Pup 10.055 0.034 10.038 0.025 10.031 RU Scl 9.252 0.025 9.286 0.020 9.285 SV Scl 10.567 0.024 10.605 0.075 10.599 AN Ser 9.895 0.026 9.895 0.020 9.914 V440 Sgr 9.104 0.025 9.218 0.031 9.230 V675 Sgr 9.088 0.025 9.114 0.022 9.059 MT Tel 8.055 0.055 8.135 0.034 8.193 AM Tuc 10.677 0.028 10.713 0.029 10.704 AB UMa 9.707 0.020 9.694 0.024 9.862 RV UMa 9.889 0.043 9.911 0.034 9.862 SX UMa 10.152 0.028 10.200 0.021 10.150 TU UMa 8.735 0.020 8.747 0.021 8.717
(This table is available in machine-readable form.)
13
Although we do not consider RRc stars here, the multiband analysis by Monson et al.(2017) finds an RRc zero-point that is consistent with Kollmeier
et al.(2013). These cuts result in a final sample of with 4, 3, and 20 RRab stars
5.1. Period–Luminosity–Metallicity Relations
In order to examine the effect of metallicity on our zero-point measurement we also consider a period–luminosity– metallicity (PLZ) relation for F160W. Our PLZ relation takes the following form:
M a log P 0.3 b c Fe H 1.60 ,
2
= ´ - + + ´ +
l l [ ( ) ] l l ([ ] )
( ) where a, b, and P take the same form as our PL relation,[Fe/H] is the metallicity of each RR Lyrae star (as compiled and homogenized by Monson et al. 2017, from Feast et al.2008; Fernley et al.1998, and Fernley & Barnes1997, see Table1.), and c is the slope of the adopted metallicity correction applied to each star, with a value of c=0.186 (Braga et al.2015). The
results are shown in Table5for comparison. For the stars with HST parallaxes, we find a fairly consistent zero-point. RR
Lyrae stars from the TGAS parallax sample exhibit a larger zero-point discrepancy, but at present the errors on the absolute magnitudes result in too large a scatter to draw a significant conclusion about this difference.
5.2. Astrometry-based Luminosity
In their analysis of RR Lyrae TGAS parallaxes, Gaia Collaboration et al. (2017) note that parallax inversion to
calculate absolute magnitudes is inadvisable if the parallax errors are too large. We mirror their analysis here for comparison and include a measurement of the zero-point using ABL(Arenou & Luri1999), defined as:
a=100.2M =v0.2mo-2. ( )3 This is the same as the Gaia Collaboration et al. (2017)
formula(2), where M is absolute magnitude, ϖ is the parallax, mo is the extinction-corrected apparent magnitude. We then solve the following PL and PLZ relations for the zero-point bλ, corresponding to Gaia Collaboration et al.(2017) formulas (4)
and(5):
100.2(al+bl)=v100.2mo-2 ( )4 100.2(al+ +bl c Fe Hl[ ])=v100.2mo-2. ( )5 We apply the samefixed slopes a and c used in Sections5.1
and5.2, the resulting zero points are given in Table5. We use a less conservative cut of σπ/π<0.2. With this less stringent cut, our usable TGAS sample includes 13 RRab stars for the ABLfit, while our HST and DR2 samples remain the same size.
5.3. Blazhko Effect
Several of the stars in our sample exhibit the Blazhko effect, which can affect the measurement of mean magnitudes in the
Table 4 Absolute Magnitudes
Name H(mag) [F160W] σ Blazhko
2MASS HST mag HST/FGS XZ Cyg −0.149 −0.275 0.221 * SU Dra −0.554 −0.543 0.245 L RR Lyr −0.537 −0.663 0.076 * UV Oct −0.588 −0.569 0.127 * TGAS XZ Cyg −0.297 −0.423 0.320 * DX Del −0.132 −0.198 0.288 L SU Dra −0.538 −0.528 0.425 L SV Eri −0.157 −0.164 0.421 L RR Gem −0.672 −0.754 0.763 * TT Lyn −0.226 −0.207 0.568 L UV Oct −0.231 −0.132 0.247 * BH Peg −0.055 −0.042 0.341 * RU Scl 0.173 0.152 0.429 L RV UMa −0.806 −0.775 0.675 * TU UMa −0.459 −0.440 0.458 L Gaia DR2 SW And −0.180 −0.163 0.200 * RR Cet −0.456 −0.495 0.112 L XZ Cyg −0.281 −0.407 0.043 * DX Del −0.099 −0.166 0.050 L SU Dra −0.582 −0.572 0.051 L SV Eri −0.881 −0.963 0.106 L RR Gem −0.569 −0.510 0.149 * TW Her −0.080 0.033 0.073 L RR Leo −0.282 −0.343 0.191 L TT Lyn −0.923 −0.882 0.068 L UV Oct −0.368 −0.349 0.034 * BH Peg −0.740 −0.641 0.088 * HK Pup −0.838 −0.808 0.128 L RU Scl −0.577 −0.599 0.148 L AN Ser −0.209 −0.236 0.105 L V440 Sgr −0.092 −0.213 0.068 L V675 Sgr −0.786 −0.757 0.115 L AB UMa −0.379 −0.345 0.070 L RV UMa −0.327 −0.296 0.078 * TU UMa −0.325 −0.307 0.090 L
(This table is available in machine-readable form.)
Table 5 Zero Points
Sample Filter Method Zero-point
Zero-point, No Blazhko HST/FGS H LSQ −0.36±0.07 −0.29±0.25 HST/FGS F160W LSQ −0.42±0.06 −0.28±0.25 HST/FGS F160W LSQ, [Fe/H] −0.42±0.07 −0.24±0.25 TGAS H LSQ −0.30±0.08 L TGAS F160W LSQ −0.30±0.09 L TGAS F160W LSQ, [Fe/H] −0.37±0.13 L DR2 H LSQ −0.39±0.04 −0.37±0.05 DR2 F160W LSQ −0.39±0.04 −0.38±0.05 DR2 F160W LSQ, [Fe/H] −0.47±0.04 −0.44±0.04 HST/FGS H ABL −0.40±0.14 −0.43±0.29 HST/FGS F160W ABL −0.39±0.14 −0.43±0.29 HST/FGS F160W ABL, [Fe/H] −0.39±0.14 −0.44±0.29 TGAS H ABL −0.39±0.08 −0.43±0.12 TGAS F160W ABL −0.39±0.08 −0.43±0.12 TGAS F160W ABL, [Fe/H] −0.38±0.08 −0.42±0.12 DR2 H ABL −0.39±0.07 −0.39±0.08 DR2 F160W ABL −0.39±0.07 −0.40±0.08 DR2 F160W ABL, [Fe/H] −0.37±0.07 −0.38±0.08
optical as well as the IR (e.g., Jurcsik et al.2018). In order to
consider the impact of the Blazhko pulsators sample, wefit PL relations with only those RRab stars that apparently do not suffer from amplitude modulation. Thefits are performed in the same manner as described above, with 1 and 14 stars from the HST and Gaia DR2 parallax samples respectively. There are no non-Blazhko stars with TGAS parallaxes below the σπ/π<0.10 cut, and 6 below the σπ/π<0.20 cut. As such, we only calculate a zero-point using the ABL method in this case. The results are given in Table5. The largest difference is for the HST parallax sample when performing a least-squares fit, as the zero-point is now constrained by a single star. The ABL method shows consistent results between the zero points fit with and without Blazhko RRab.
6. Conclusions
We have carried out photometry using HST/WFC3 of 30 Galactic RR Lyrae stars taken in F160W. Our results are part of the larger effort associated with the CCHP; the photometry presented here will be used to provide a calibration of the Galactic RR Lyrae zero-point in the same flight-magnitude system used to measure RR Lyrae distances to nearby galaxies. The mean magnitudes that we calculate are consistent with ground-based H-band photometry, within the expected offset between the 2MASS and HST flight-magnitude systems.
We also present PL relations for the RRab stars in our sample using both HST/FGS and Gaia TGAS & DR2 parallaxes. These PL relations provide an initial estimate of the HST F160W zero-point, with all three parallax samples providing zero points consistent within their respective errors. Our final Galactic RR Lyrae zero-point will rely on a larger sample of more accurate parallaxes provided by the next Gaia data release.
We thank the referee for their helpful and constructive comments. M.G.L. was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) (No. 2017R1A2B4004632). This paper makes use of data from the European Space Agency (ESA) mission Gaia(https://www.cosmos.esa.int/gaia), processed
by the Gaia Data Processing and Analysis Consortium (DPAC,https://www.cosmos.esa.int/web/gaia/dpac/
consortium). Funding for the DPAC has been provided by
national institutions, in particular, the institutions participating in the Gaia Multilateral Agreement.
ORCID iDs
Jeffrey A. Rich https://orcid.org/0000-0002-5807-5078
Barry F. Madore https://orcid.org/0000-0002-1576-1676
Rachael L. Beaton https://orcid.org/0000-0002-1691-8217
Dylan Hatt https://orcid.org/0000-0003-2767-2379
Figure 3.Period–luminosity (PL) relations for our Galactic RR Lyrae; only RRab (fundamental pulsators) are used. The left panel consists of the four RRab with HST/FGS parallaxes from Benedict et al. (2011), the right panel shows RRab in the Gaia DR2 parallax sample, our F160W sample contains 20 stars (see Section5for a discussion of our selection method). We compare our F160W PL relation with a fit to ground-based 2MASS H-band magnitudes (Monson et al.2017).
Taylor Hoyt https://orcid.org/0000-0001-9664-0560
In-Sung Jang https://orcid.org/0000-0002-2502-0070
Juna A. Kollmeier https://orcid.org/0000-0001-9852-1610
Myung Gyoon Lee https://orcid.org/0000-0003-2713-6744
Jillian R. Neeley https://orcid.org/0000-0002-8894-836X
Victoria Scowcroft https://orcid.org/0000-0001-8829-4653
Mark Seibert https://orcid.org/0000-0002-1143-5515
References
Arenou, F., & Luri, X. 1999, in ASP Conf. Ser. 167, Harmonizing Cosmic Distance Scales in a Post-HIPPARCOS Era, ed. D. Egret & A. Heck(San Francisco, CA: ASP),13
Astraatmadja, T. L., & Bailer-Jones, C. A. L. 2016,ApJ,832, 137
Beaton, R. L., Freedman, W. L., Madore, B. F., et al. 2016,ApJ,832, 210
Benedict, G. F., McArthur, B. E., Feast, M. W., et al. 2011,AJ,142, 187
Bono, G., Caputo, F., Castellani, V., et al. 2003,MNRAS,344, 1097
Braga, V. F., Dall’Ora, M., Bono, G., et al. 2015,ApJ,799, 165
Brown, T. M., Casertano, S., Strader, J., et al. 2018,ApJL,856, L6
Casertano, S., Riess, A. G., Anderson, J., et al. 2016,ApJ,825, 11
Catelan, M., Pritzl, B. J., & Smith, H. A. 2004,ApJS,154, 633
Clementini, G. 2016, CoKon,105, 3
Dambis, A. K., Rastorguev, A. S., & Zabolotskikh, M. V. 2014, MNRAS,
439, 3765
de Bruijne, J. H. J., Rygl, K. L. J., & Antoja, T. 2014, in EAS Publications Ser. 67, The Milky Way Unravelled by Gaia: GREAT Science from the Gaia Data Releases, ed. N. A. Walton et al.(Paris: EDP Sciences),23
Feast, M. W., Laney, C. D., Kinman, T. D., van Leeuwen, F., & Whitelock, P. A. 2008,MNRAS,386, 2115
Fernley, J., & Barnes, T. G. 1997,A&AS,125, 313
Fernley, J., Barnes, T. G., Skillen, I., et al. 1998, A&A,330, 515
Freedman, W. L. 2017, AAS Meeting, 229,234.01
Freedman, W. L., & Madore, B. F. 2010a,ApJ,719, 335
Freedman, W. L., & Madore, B. F. 2010b,ARA&A,48, 673
Freedman, W. L., Madore, B. F., Scowcroft, V., et al. 2012,ApJ,758, 24
Freedman, W. L., Madore, B. F., Gibson, B. K., et al. 2001,ApJ,553, 47
Gaia Collaboration, Clementini, G., Eyer, L., et al. 2017,A&A,605, A79
Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018,A&A,616, A1
Gould, A., & Kollmeier, J. A. 2017,JKAS,50, 1
Hubble, E. 1929,PNAS,15, 168
Jurcsik, J., Hajdu, G., Dékány, I., et al. 2018,MNRAS,475, 4208
Kollmeier, J. A., Szczygieł, D. M., Burns, C. R., et al. 2013,ApJ,775, 57
Lindegren, L., Lammers, U., Bastian, U., et al. 2016,A&A,595, A4
Lindegren, L., Hernández, J., Bombrun, A., et al. 2018,A&A,616, A2
Longmore, A. J., Dixon, R., Skillen, I., Jameson, R. F., & Fernley, J. A. 1990, MNRAS,247, 684
Longmore, A. J., Fernley, J. A., & Jameson, R. F. 1986,MNRAS,220, 279
Madore, B. F., & Freedman, W. L. 2012,ApJ,744, 132
Madore, B. F., Hoffman, D., Freedman, W. L., et al. 2013,ApJ,776, 135
Monson, A. J., Beaton, R. L., Scowcroft, V., et al. 2017,AJ,153, 96
Neeley, J. R., Marengo, M., Bono, G., et al. 2015,ApJ,808, 11
Persson, S. E., Madore, B. F., Krzemiński, W., et al. 2004,AJ,128, 2239
Riess, A. G. 2011, An Independent Determination of WFC3-IR Zeropoints and Count Rate Non-Linearity from 2MASS Asterisms, Tech. rep. 2011-15
Riess, A. G., Casertano, S., Anderson, J., MacKenty, J., & Filippenko, A. V. 2014,ApJ,785, 161
Riess, A. G., Macri, L. M., Hoffmann, S. L., et al. 2016,ApJ,826, 56
Riess, A. G., Casertano, S., Yuan, W., et al. 2018,ApJ,855, 136
Stetson, P. B. 1987,PASP,99, 191
Stetson, P. B. 1990,PASP,102, 932
Stetson, P. B. 1994,PASP,106, 250
